Subject: Re: [sv-ac] R29a/b Optional or mandatory name for assertions/prop erties/assumptions.
From: Prakash Narain (prakash@realintent.com)
Date: Wed Sep 18 2002 - 15:03:33 PDT
Actually I am looking for you to make that call in your context
at Sun.
Suppose your design has 10,000 assertions. If you ask your
50 designers to enter a name for each assertion in their modules
will they balk arguing too much typing?
Suppose you are using multiple tools/methodologies and want
to convey some information to the simulator based upon Formal
Analysis or to Formal Analysis based upon simulation. Would
you like to continuously update this communication file because
names keep on changing? Suppose you are just doing Formal
Analysis and are making use of name based tool controls. Would
you like to continuously update the script because names keep
changing?
Finally how would you ensure that your simulator and other tools
have a consistent naming mechanism for assertions.
So what will be your position? Here are some options:
1) You will not mandate naming of assertions.
2) You will mandate naming of assertions but would
prefer to enforce that by creating a flow based
upon lint tools that will require somebody to police
the process.
3) You prefer that the language requires that each assertion
be named.
Best Regards,
Prakash
Shrenik Mehta wrote:
>.Date: Wed, 18 Sep 2002 13:39:28 -0700
>.From: Prakash Narain <prakash@realintent.com>
>.
>.However, the harder we make for engineers to debug feedback from the tool
>.and to maintain their tool control scripts while their design is
>.changing, the
>.harder it will be to get the assertions adopted. Today I analyzed ovi1751
>.and fixed a bug. Tomorrow ovi1769 requires debugging.
>.
>.Let us use a different approach.
>.
>.I say that if I am a design manager using assertions I will make naming of
>.assertions mandatory. Let us hear from other design manager wannabes
>.out their. Gail, what would you do?
>
>
>So on a project, if we have to write about 10,000 assertions how would it
>work or not work?
>
>I have not followed the thread too carefully, so I might have missed the
>context.
>
>
>Shrenik
>.
>.Best Regards,
>.
>.Prakash
>.
>.Tom Anderson wrote:
>.
>.>Well said, Gail. One additional point is that we want to encourage
>.>engineers to use assertions. The more characters we require for each
>.>assertion, the harder we make it look.
>.>
>.>Tom A. [aka ovi1751]
>.>
>.>.
>.>
>.
>.
>
>---------------------------------------------------------------------------
>Shrenik Mehta Tel: 408-774-8580
>Senior Engineering Manager Fax: 408-774-8154
>Global Testability, Tools and Validation
>Computer Systems email:shrenik.mehta@Sun.Com
>Sun Microsystems pager:shrenik.mehta@pager.Eng.Sun.COM
>---------------------------------------------------------------------------
>
>
>
This archive was generated by hypermail 2b28 : Wed Sep 18 2002 - 15:05:20 PDT